Re: si_band type

From: Jakub Jelinek (jakub@redhat.com)
Date: Thu Feb 24 2000 - 09:33:36 EST


On Thu, Feb 24, 2000 at 12:35:12PM +0000, Stephen C. Tweedie wrote:
> Hi,
>
> On Wed, 23 Feb 2000 09:35:11 +0100, Jakub Jelinek <jakub@redhat.com>
> said:
>
> > It should be doable for sparc64, I guess ia64 as well (but as I don't know
> > if some ia64 application already uses si_band so I'm not enclosing a patch
> > for it,
>
> si_fd is available in the 2.2 kernel ABI, so if you want to move si_band
> I'd prefer to avoid relocating si_fd at the same time.
>
> > /* SIGPOLL */
> > struct {
> > - int _band; /* POLL_IN, POLL_OUT, POLL_MSG */
> > + long _band; /* POLL_IN, POLL_OUT, POLL_MSG */
> > int _fd;
> > } _sigpoll;
> > } _sifields;
>
> What about
>
> {
> int _padding; /* Old location of int _band */
> int _fd;
> long _band; /* POLL_IN, POLL_OUT, POLL_MSG */
> };
>
> as a solution?

That was a sparc64 patch, and sparc 64bit userland programs are not
supported until some 2.3.40 or so (and even then it's me and at most 2-3
other people who run it so far).
The kernel translates it into int _fd; int _band; anyway during copy to
userland for 32bit apps (but it is imperfect and unless my siginfo patch
makes it in or some other similar solution is added, it will be always).

Cheers,
    Jakub
___________________________________________________________________
Jakub Jelinek | jakub@redhat.com | http://sunsite.mff.cuni.cz/~jj
Linux version 2.3.47 on a sparc64 machine (1343.49 BogoMips)
___________________________________________________________________

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Feb 29 2000 - 21:00:10 EST